The pharmaceutical industry now recognises the importance of the newly defined discipline of pharmacometrics. Pharmacometrics uses mathematical models to describe and then predict the performance of new drugs in clinical development. To ensure these models are useful, the clinical studies need to be designed such that the data generated allows the model predictions to be sufficiently accurate and precise. The capability of the available software to reliably estimate the model parameters must also be well understood.  This thesis investigated two important areas in pharmacometrics: optimal design and software estimation performance. The three optimal design papers progressed significant areas of optimal design research, especially relevant to phase II dose response designs. The use of exposure, rather than dose, was investigated within an optimal design framework. In addition to using both optimal design and clinical trial simulation, this work employed a wide range of metrics for assessing design performance, and was illustrative of how optimal designs for exposure response models may yield dose selections quite different to those based on standard dose response models. The investigation of the optimal designs for Poisson dose response models demonstrated a novel mathematical approach to the necessary matrix calculations for non-linear mixed effects models. Finally, the enormous potential of using optimal adaptive designs over fixed optimal designs was demonstrated. The results showed how the adaptive designs were robust to initial parameter misspecification, with the capability to "learn" the true dose response using the accruing subject data. The two estimation performance papers investigated the relative performance of a number of different algorithms and software programs for two complex pharmacometric models. In conclusion these papers, in combination, cover a wide spectrum of study designs for non-linear dose/exposure response models, covering: normal/non-normal data, fixed/mixed effect models, single/multiple design criteria metrics, optimal design/clinical trial simulation, and adaptive/fixed designs.

ABSTRACT Occupational exposure to wood dust and biohazards associated with wood dust (endotoxins, (1->3)-b-D-glucans, Gram (-)ve bacteria and fungi), their correlation to respiratory function, and symptoms among woodworkers have been investigated in the present study. Wood dust, endotoxins, and allergenic fungi are the main hazards found in woodworking environments. Relatively very few studies have been done on wood dust exposure. The present study was designed to comprehensively investigate the health effects of wood dust exposure, and in particular provide new information regarding: Exposure to (1->3)-b-D-glucans in an occupational environment; Levels of exposure to wood dust and biohazards associated with wood dust in different woodworking environments; Correlations among personal exposures, especially correlations between (1->3)-b-D-glucans and fungi exposures, and endotoxins and Gram (-)ve bacteria exposures; Effects of personal exposure to biohazards on lung function; Effects of personal exposure to biohazards on work-related symptoms; and Determinants of inhalable exposures (provide which factors in the environment influence the personal inhalable exposures). Workers at four different woodworking processes; two logging sites, four sawmills, one major woodchipping operation and five joineries situated in the state of New South Wales in Australia were studied for personal exposure to inhalable dust (n=182) and respirable dust (n=81), fungi (n=120), Gram (-)ve bacteria (n=120), inhalable endotoxin (n=160), respirable endotoxin (n=79), inhalable (1->3)-b-D-glucan (n=105), and respirable (1->3)-b-D-glucan (n=62). The workers (n=168) were also tested for lung function. A questionnaire study (n=195) was carried out to determine the prevalence of work-related symptoms. The geometric mean inhalable exposure at logging sites was 0.56 mg/m3 (n=7), sawmills 1.59 mg/m3 (n=93), the woodchipping mill 1.86 mg/m3 (n=9) and joineries 3.68 mg/m3 (n=66). Overall, sixty two percent of the exposures exceeded the current standards. Among joineries, 95% of the hardwood exposures and 35% of the softwood exposures were above the relevant standards. Compared with green mills, the percentage of samples, which exceeded the hardwood standard was high for dry mills (70% in dry mills, 50% in green mills). The respirable dust exposures were high at the joineries compared with the other worksites. Exposure levels to fungi at logging sites and sawmills were in the range 103-104 cfu/m3, woodchipping 103-105 cfu/m3 and joineries 102-104 cfu/m3. The predominant fungi found at sawmills were Penicillium spp. High exposure levels of Aureobasidium pullulans were also found at two sawmills. At the woodchipping mill the predominant species were Aspergillus fumigatus, Penicillium spp., and Paecilomyces spp. The sawmills, which employed kiln drying processes, had lower exposure levels of fungi compared with the green mills. Those workplaces which had efficient dust control systems showed less exposure to fungi and bacteria. Although mean endotoxin levels were lower than the suggested threshold value of 20 ng/m3, some personal exposures at sawmills and joineries exceeded the threshold limit value. The mean inhalable (1->3)-b-D-glucan level at the woodchipping mill was 2.32 ng/m3, at sawmills 1.37 ng/m3, at logging sites 2.02 ng/m3, and at joineries 0.43 ng/m3. For the respirable size fraction, mean endotoxin and mean (1->3)-b-D-glucan concentrations were much lower, being similar to observed dust concentrations. Significant correlations were found between mean inhalable endotoxin and Gram (-)ve bacteria levels (p<0.0001), and mean airborne inhalable (1->3)-b-D-glucan and fungi levels (p=0.0003). The correlations between mean respirable endotoxin levels vs Gram (-)ve bacteria exposure levels (p=0.005), and respirable (1->3)-b-D-glucan exposure levels vs total fungi levels (p=0.005) were also significant. Significant correlations were found between lung function and personal exposures. Multivariate analyses showed that the effect of all the personal exposures on cross-shift decrements in lung function was more prominent among sawmill and chip mill workers compared with joinery workers. Woodworkers had markedly high prevalence of cough, phlegm, chronic bronchitis, frequent headaches, throat and eye irritations, and nasal symptoms compared with controls. Among the woodworkers, smokers had a high prevalence of chronic bronchitis (20%) compared with non-smokers (10%). Some workers also reported a variety of allergy problems due to exposure to various types of wood dust. Both joinery workers and sawmill and chip mill workers revealed significant correlations between work-related symptoms and personal exposures. Chronic bronchitis was significantly correlated with personal exposure to wood dust, endotoxin, (1->3)-b-D-glucan, fungi, and Gram (-)ve bacteria among joinery workers. Whereas among sawmill workers chronic bronchitis was significantly correlated with personal exposure to endotoxin, (1->3)-b-D-glucan, and fungi. Woodworkers showed significant positive correlations between percentage cross-shift change (decrease) in lung function and respiratory symptoms. Significant inverse correlations were also found among percentage predicted lung function and respiratory symptoms. The elevated inhalable dust exposures observed in this study can be explained by a combination of factors, including: lack of awareness of potential health effects of wood dust exposure among both management and workers, aging equipment, inadequate and ineffective dust extraction systems or usually none especially for hand held tools, poor maintenance of the ventilation system in some, non-segregation of dusty processes, dry sweeping, and the use of compressed air jets. The determinant-of-exposure analysis confirmed the field observations. The significant determinants of personal inhalable dust exposures (n=163) were found to be: local exhaust ventilation, job title, use of hand-held tools, cleaning method used, use of compressed air, and green or dry wood processed. Type of wood processed was not found to be statistically significant. A majority of workers (~90%) did not wear appropriate respirators approved for wood dust, while the workers who did wear them, used them on average less than 50% of the time. Workers should be protected by controlling dust at its source. When exposure to wood dust cannot be avoided, engineering controls should be supplemented with the use of appropriate personal protective equipment.

Diss. (sammanfattning) Stockholm : Karolinska institutet, 2005. / Härtill 5 uppsatser.

Diss. (sammanfattning) Stockholm : Karolinska institutet, 2005. / Härtill 5 uppsatser.

Diss. (sammanfattning) Linköping : Univ., 2003. / Härtill 6 uppsatser.

Diss. (sammanfattning) Stockholm : Karolinska institutet, 2005. / Härtill 5 uppsatser.

Thesis--University of Oklahoma. / Bibliography: leaves 81-83.

Thesis--University of Oklahoma. / Bibliography: leaves 78-79.

Thesis (M.Sc.)--Memorial University of Newfoundland, 2004. / Bibliography: leaves 94-103.

Ο βασικός στόχος αυτής της εργασίας είναι η μελέτη της θεωρητικής συμπεριφοράς και τα πλεονεκτήματα της γενικευμένης βάθμωσης δόσης-απόκρισης, καθώς και η έρευνα για τη χρησιμότητα της γενικευμένης βάθμωσης δόσης-απόκρισης σε πρακτικά ακτινοβιολογικά πλάνα θεραπείας μέσω χρήσης της πλατφόρμας RayStation. Τέλος, θα διερευνηθεί η επίδραση της αρχιτεκτονικής του οργάνου (παράλληλη / σειριακή). Βασικό υλικό της μελέτης μας ήταν το λογισμικό για το σχεδιασμό πλάνων θεραπείας RAYSTATION 1.9, που αναπτύχθηκε και σχεδιάστηκε από τη RAYSEARCH LABORATORIES AB, Στοκχόλμη, Σουηδία. Εκτός από αυτό, κάναμε εκτεταμένη χρήση βασικών θεωρητικών τύπων που σχετίζονται με τον υπολογισμό του TCP, NTCP, P + και του γ(D) καθώς και με την ακτινοβιολογικά μοντέλα. Σε ό, τι αφορά τις μεθόδους της μελέτης αυτής, πρώτον υπολογίσαμε θεωρητικά το γ(D) για TCP και NTCP αντίστοιχα, για την ετερογενή κατανομή της δόσης σε διαφορετικά μεγέθη, προκειμένου να επαληθεύσουμε τους υπολογισμούς του RAYSTATION για TCP και NTCP. Επιπλέον, έχουμε δημιουργήσει ένα πλάνο θεραπείας με το όργανο-στόχος και τα όργανα που βρίσκονται σε κίνδυνο να βρίσκονται στην ίδια περιοχή ενδιαφέροντος, προκειμένου να ελέγξουμε την εγκυρότητα του συστήματος για την συνάρτηση P + καθώς και των γενικευμένων γ(D). Επιπλέον, έχουμε θέσει μια σειρά από διαφορετικά πλάνα θεραπείας με το όργανο-στόχος και τα όργανα σε κίνδυνο σε διαφορετικές περιοχές ενδιαφέροντος όπου αυξήσαμε τη μέση δόση, προκειμένου να διερευνήσουμε τη συμπεριφορά του ΔP(μεταβολή απόκρισης) και του γ(D), πριν και μετά την αλλαγή της δοσολογίας. Επίσης υπολογίσαμε θεωρητικά τις ποσότητες αυτές, προκειμένου να εξακριβωθεί η εγκυρότητα των θεωρητικών εκφράσεων συγκρίνοντας τες με τις τιμές που το σύστημα παρήγε σε μας. Τέλος, προσπαθήσαμε να διερευνήσουμε τη συμπεριφορά της ποσότητας ΔP υπολογίζοντας το σχετικό σφάλμα μεταξύ της πραγματικής και την κατά προσέγγιση τιμής χρησιμοποιώντας το Poisson και το Probit μοντέλο, για την περίπτωση όπου έχουμε ένα όργανο-στόχος το οποίο αποτελείται από δύο τμήματα σε παράλληλη αρχιτεκτονική και με τον ίδιο αριθμό κλώνων. Όσον αφορά τα αποτελέσματά μας, πρώτα απ 'όλα, επαληθεύσαμε θεωρητικά τους υπολογισμούς του RAYSTATION για τo γενικευμένο γ(D) και την αντικειμενική συνάρτηση με τη χρήση ενός ανεξάρτητου τρόπου υπολογισμών. Επιπλέον, αποδείχθηκε ότι μετά από μια μικρή μεταβολή (αύξηση) της δόσης, το όργανο που έχει επηρεαστεί περισσότερο, είναι το όργανο με το υψηλότερο γενικευμένο γ(D). Εκτός από αυτό, ελέγχθηκε η εγκυρότητα των θεωρητικών εκφράσεων σχετικά με τον υπολογισμό της μεταβολής της απόκρισης και του γενικευμένου γ(D), αλλά μόνο για την περίπτωση μικρής μεταβολής της δόσης. Ειδικά για την περίπτωση του 50% TCP και NTCP, οι θεωρητικές τιμές που το σύστημα παρέχει εμφανίζουν μεγάλη προσέγγιση με τις πειραματικές, γεγονός που αποδεικνύει τη μεγάλη σημασία του D50 μοντέλου στο προσδιορισμό των κλινικών επιπέδων απόκρισης. Τέλος, όσον αφορά το τελευταίο μέρος των υπολογισμών μας, μπορούμε εύκολα να πούμε ότι η συμπεριφορά της ΔPapprox εμφανίζεται λογική, διότι, για τα δύο μοντέλα που χρησιμοποιήσαμε, πλησιάζει σημαντικά την πραγματική ΔP γύρω από την περιοχή του 50% ή 37%, όπως και αναμέναμε. Επαληθεύσαμε σε αρκετά ικανοποιητικό επίπεδο κάποιες βασικές θεωρητικές προσεγγίσεις για την κλίση δόσης-απόκρισης σχετικά με τη μη ομοιόμορφη κατανομή δόσης μέσω της πλατφόρμας RayStation αλλά το πιο σημαντικό πράγμα είναι το γεγονός ότι η χρησιμότητα της των γενικευμένης βάθμωσης δόσης-απόκρισης είναι εξαιρετικά σημαντική, διότι δίνει στο σχεδιαστή των πλάνων θεραπείας τη δυνατότητα να ερευνήσει ακριβώς το όργανο το οποίο, θα επηρεαστεί περισσότερο μετά από μια μικρή αύξηση της δόσης και ως εκ τούτου θα είναι σε θέση να βελτιστοποιήσει το πλάνο για αύξηση ελέγχου του όγκου αλλά και ελαχιστοποίηση επιπλοκών των υγειών ιστών. / The basic aim of that work is the study of the theoretical behavior and merits of the Generalized Dose-Response gradient as well as the investigation of the usefulness of the generalized dose response gradient in practical radiobiological treatment planning through the use of RayStation. Last but not least, it will be investigated the influence of the organ architecture(parallel/serial). Basic material of our study was the treatment planning platform RAYSTATION 1.9 that was developed and designed by RAYSEARCH LABORATORIES AB,STOCKHOLM,SWEDEN. Except for that ,we made extensive use of basic theoretical formulas that are related to the calculation of TCP, NTCP, P+ and Generalized Gamma as well as to the radiobiological models. As far as the methods of that study are concerned, firstly we calculated theoretically the Generalized Gamma for TCP and NTCP respectively, for heterogeneous dose distribution to different volumes in order to verify RAYSTATION computations for TCP and NTCP. Furthermore, we set a treatment plan with the target organ and the organs at risk in the same ROI in order to check the validity of the system concerning the objective function P+ and the Generalized Gamma. Moreover ,we set a number of different treatment plans with the target organ and the organs at risk in different ROIs and we increased the mean dose in order to investigate the behavior of change in response and that of γ(D) ,before and after the change in dose and to calculate theoretically these quantities, in order to verify the validity of the theoretical expressions by comparing them with the values that the system is providing to us. Finally, we tried to investigate the behavior of ΔP by calculating the relative error between the real and the approximate value using the Poisson and the Probit model, for the case of having a target organ consisting of two compartments in a parallel architecture and with the same number of clonogens. Concerning our results, first of all, we verified theoretically the computations of the RAYSTATION about the Generalized Gamma and the objective function by using an independent way of calculations. Furthermore, we proved that after a small change (increase) in dose ,the organ that is being affected most ,is the organ with the highest Generalized Gamma. Except for that, we verified the validity of the theoretical expressions concerning the calculation of the change in response and that of Generalized Gamma but only for the case of small change in dose. Especially for the case of 50% TCP and NTCP, the theoretical and the values that the system is providing appear great approximation, a fact that proves the high importance of D50 model in specifying clinical response levels. Finally, concerning the last part of our calculations, we easily can say that the behavior of ΔPapprox looks sensible because, for both models that we used, it approaches significantly the real ΔP around the region of 50% or 37% response, as we were expecting. We verified in a quite satisfying level some basic theoretical approaches for dose-response gradient concerning the non-uniform dose delivery through the RayStation platform but the most important thing is the fact that the usefulness of the of the Generalized Dose response gradient is extremely important because it gives to the planner the opportunity to investigate precisely which organ, from the normal tissues will be affected most after a small increase in dose and as a result he will be able to optimize the plan for higher tumor control and lowest normal tissue complications. *This work had been done in collaboration with the Division of Medical Radiation Physics, Department of Oncology-Pathology, Karolinska Institute, Stockholm, Sweden